Total Cost of Ownership in PV Manufacturing Guide
Publication date: October 2013
I. Executive Summary………......….….1
II. SEMI Standards…………….………2
III. VDMA………………………….…6
IV. Comparison VDMA and SEMI......10
V. Conclusion…………………….….12
VI. References………………………..12
NOTICE:
This guide is intended only to provide an
overview of the benefits of strategy-oriented
cost management in the PV industry through
the use of cost of ownership (COO) and/or
lifecycle cost (LCC) calculations and analyses
for better decision making. To help the reader
to better understand these cost calculations
and their associated standards, it also contains
a high-level overview and comparison of
some key corresponding SEMI and VDMA
Standards related to measuring equipment
performance for reliability, availability,
maintainability (RAM), and productivity that
are key inputs into these cost analyses. This
guide is not intended and should not
replace the use of the actual standards
that contain the necessary details for correct
implementation and compliance.
Jointly published by SEMI and VDMA
Copyright © 2013 Semiconductor Equipment and
Materials International (SEMI) and VDMA. SEMI
and semi.org are trademarks or registered
trademarks of SEMI in the United States and other
countries. VDMA and vdma.org/vdma.de are
trademarks of VDMA. All other company and
product names are trademarks of their respective
holders.
(Revised February 2014)
I. Executive Summary
Purchasing decisions for capital goods (i.e. machinery, equipment,
components and plants) are influenced more and more by indirect
costs. Many companies and industry associations have begun to deal
with life-cycle models, such as total cost of ownership (COO) and life
cycle costs as part of their strategy-oriented cost management. The
individual models are often developed for very specific applications
and are thus difficult to compare for suppliers and buyers of industrial
products. To counter this deficiency, the usage of standards has to be
more widely accepted.
In sales negotiations throughout the photovoltaic value chain,
equipment suppliers and plant manufacturing companies are invariably
faced with the fact that customers apply simple acquisition cost
comparisons and do not thoroughly analyze the determinants for the
profitability of the individual offers. Such a comprehensive analysis
would have to include the investment project with all its direct and
indirect financial implications across the entire lifecycle. In most cases,
mechanical engineering manufacturers are not in the position to
provide a corresponding analysis, since they do not possess the
required information about the customer’s planned production
programs.
However, considering that acquisition costs alone only contribute to a
fraction of the total costs that accrue across the entire lifetime of
investment goods, these operating costs and their proportion to the
acquisition costs have to be taken into account.
In the past, the PV industry did not use existing standards in their sales
negotiations and/or discussions among suppliers and manufacturers.
After numerous discussions with stakeholders along the supply chain it
became clear that the calculation of cost of ownership is perceived as
important, however, frequently the tools are not known. As a result
of a survey among key contributors to standardization activities it was
determined that the different terminology and methodology is a key
burden for the calculation.
SEMI, the global trade association for equipment and material
suppliers, and VDMA, the German engineering association for
machinery, compared the existing standards and illustrate in this guide
how different timing views/terminology must be understood to reach
comparable results – regardless of which standard has been used.
Page |1
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

II. SEMI Standards

In photovoltaic related industries or in silicon-based technology, such as the semiconductor industry, the SEMI E10, SEMI
E35 and SEMI E79 standards* are applied to calculate the cost of ownership for equipment.

The following chapter provides an extract from SEMI E10 and E35 to better explain the content of the SEMI standards
documents.

The SEMI standard E10 establishes a common basis for communication between users and suppliers of semiconductor
manufacturing equipment by providing a standardized methodology for measuring reliability, availability, and maintainability
(RAM) and utilization performance of equipment in a manufacturing environment.

NONSCHEDULED
TIME

UNSCHEDULED
DOWNTIME
Equipment Total Time
Downtime
SCHEDULED
DOWNTIME

ENGINEERING Operations
TIME Time
Equipment
Uptime
STANDBY
TIME
Manufacturing
Time
PRODUCTIVE
TIME

Figure 1: Extract from SEMI E10 Equipment State Stack Chart

Total Time

Nonscheduled Operations
Time Time

 Unworked time periods

 Off-line training
 Installation, modification,
rebuild, or upgrade Uptime Downtime
 Shutdown/Start-up

Engineering Manufacturing Unscheduled Scheduled

Time Time Downtime Downtime

 Process engineering  Repair  Preventive maintenance

 Equipment engineering  Maintenance delay  Setup
 Software engineering  Out-of-spec input  Maintenance delay
Productive Standby
 Change of consumable  Production test
Time Time
material  Change of consumable
 Regular production  No operator  Facilities related material
 Work for third party  No units  Unspecified  Facilities related
 Rework  No support tool  Unspecified
 Manufacturing support  Associated cluster
operations tool equipment
 Engineering runs with module down
production units  No production test
 Production recipe results available
specified times  No input from ext.
automation system
 Waiting/inactive
times

Figure 2: Extract from SEMI E10 Equipment State Hierarchy

Page |2
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE
To measure equipment performance (e.g., RAM metrics),
SEMI E10 defines six basic equipment states into which all
equipment conditions and periods of time must fall. Each
equipment system must be in one and only one SEMI E10
state at any point in time. Each equipment system must be
subject to at most one failure at any point in time regardless of
the number of underlying constituent problems contributing to
or arising from that failure.
The equipment states are determined by function, not by
organization. Any given maintenance procedure, for example,
is classified the same way no matter who performs it (e.g., an
operator, a production technician, a maintenance technician, a
process engineer).
Figure 1 is a stack chart of the six basic equipment states. These
basic equipment states can be divided into as many substates as
are required to achieve the equipment tracking resolution that
a manufacturing operation desires. SEMI E10 makes no
attempt to list all possible substates, but does give some that are
required to support certain metrics and other examples for
guidance.
Key blocks of time associated with the basic states and substates
are given in Figure 2. These blocks of time are used in the
RAM equations given later in the standard SEMI E10. The
blocks of time associated with the basic states and substates are
described in the following sections.
The main purpose of Fig. 1 is to illustrate the six (and that
there are only six) basic (i.e., top level, major), mutually
exclusive equipment states that cover total time during an
observation period for each equipment system. Combinations
of the times in these six basic equipment states are then defined
into categories of times that somewhat describe how the
equipment system is being used. For example, there are two
types of downtime, scheduled (i.e., planned) and unscheduled
(i.e., unplanned), that summed together define all of the
downtime for the equipment system. Most of times in these six
basic equipment states are directly used in some of the
calculated metrics. It is important to note that the equipment
system being measured can only be in one and only one
equipment state or substate within an equipment state at any
point of time during the observation period. Also, E10 has the
flexibility to allow a user to define additional E10-compliant
metrics based on these states and substates that they may find
useful for them. For example, the user may choose to further
subdivide an individual substate into more detailed lower-level
substates to better understand the factors driving the time in
that substate and as a basis to define additional metrics.
Fig. 2 shows the same information as Fig. 1, but in a
hierarchical manner where the six basic equipment states are in
gray boxes. Under each of these six basic equipment states,
there are examples of the types of substates that exist for each
one. Some of these individual substates are also used in some of
the calculated metrics in section 8. For example, the times in
three of the scheduled downtime (SDT) substates (i.e., SDT
preventive maintenance time, SDT setup time, SDT change of
consumable material time) are summed to calculate the
equipment-dependent scheduled downtime metric, which is
then used in another calculation to determine the percentage of
time that an equipment system is in the equipment-dependent
scheduled downtime state relative to the equipment-dependent
portion of operations time.
The metrics in E10 are basically time-based metrics determined
from the times that the equipment system is in one or more of
the six basic equipment states or their individual substates.
However, there are other values used in the calculations that
are based on what happens while the equipment is in one of
the six basic equipment states or their individual substates. For
example, the number of equipment failures that occur during
the productive state is used in multiple calculations/metrics.
Most of E10 is used to carefully define how the user
determines which equipment state and substate the equipment
system being measured is in at any point in time during the
observation period as well as defining the most commonly used
and useful metrics for determining the equipment system's
reliability, availability, maintainability, and utilization
performance. These states/times are also used in other SEMI
Standards, such as in E79 as the basis to determine other
equipment system's performance metrics related to its
productivity.
This latest version of E10 also deals with the more complicated
equipment systems called multi-path cluster tools where
different portions (e.g., modules) of the equipment system may
be in different states and substates during the same point in
time. E10 now provides an official, systematic way to
determine the overall equipment system's one state or substate
at any point in time based on the combination of the often
different states or substates that its individual portions are in at
that same point in time. Note that the E10 metrics can be
separately determined for each of these individual portions of
the equipment system with very different values. For example,
one particular module could be 100% down for an entire
observation period while the multi-path cluster tool it is a part
of could be 100% up and available for running production
during that same observation period.
Page |3
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE
Figure 3: Extract from SEMI E35
Relationship of Factors to Difficulty of Collection and Validation
Certain cost factors are more difficult than others to accurately
determine. Figure 3 depicts the relationship of some of the
COO model input factors in order to have the ability to collect
and validate them. The accuracy of a COO calculation may
be prone to a variety of errors or omissions.
The 'cost factor pyramid' is intended to break out many of the
individual cost factors and show a relative relationship between
them based on the typical level of difficulty in identifying and
accurately determining the values for them to include in the
overall COO calculation. These individual cost factors basically
relate the individual cost elements of each cost category to each
other based on this level of difficulty in accurately determining
their values. For example, the purchase price of the equipment
is generally the easiest direct cost factor to obtain and include
in the calculation while accurate defect limited yield and
parametric limited yield values are typically the hardest ones.
This is because it is very difficult to accurately determine the
specific impact of individual equipment and specific process(es)
to these ‘end of the process flow’ electrical measurements that
are impacted by all of the individual equipment and
their processes in the production process. These yield cost
factors can and often will 'overwhelm' the impact of all of the
other cost factors and inaccuracies in them can cause the
overall accuracy of the COO value to be greatly reduced. For
this reason, the COO is often calculated as separate values with
and without these factors included to better understand them
and their impact on the overall COO analysis.
Bottlenecks, cycle times, and line balance interrelationships are
not included in the COO calculation. They are not included
because it is difficult, in an individual COO model, to show
the impact of equipment being modeled on the complete
manufacturing factory. A factory-level cost and/or simulation
model should be used for these purposes.
The interrelationship cost factors are shaded to show that they
are different from the others in the pyramid. As explained,
there is no attempt to include these real cost factors in the
COO calculation as defined later in E35 as they are much
beyond the scope and require a much more comprehensive
cost model with orders of magnitude more data needed to be
able to include them. Typically, these more comprehensive
cost models do not provide the granularity in understanding of
the individual cost factors, nor the impact of a good COO
model. While the COO value calculated is considered to
be 'incomplete' by omitting them, it can still be extremely
useful for the specific purpose of the analysis for which it is
being used. The COO modeler needs to understand this
limitation of what can and cannot be included when making
decisions based on a COO analysis and whether it is
appropriate to use a COO model to accurately address the
question(s) to be answered for the purpose of the analysis.
A COO calculation may have more detail than presented
explicitly in this SEMI E35 Guide. The structure of the
Guide, however, allows for these situations.
It is virtually impossible to include every possible cost factor of
each individual cost element in their descriptions. Several of
the cost factors can be further subdivided into additional inputs
and modeled. For example, burdened labor costs for each
labor type is usually an overall estimate that includes the direct
hourly wage plus burdened cost factors such as benefits, taxes,
etc. While these actually vary by individual employee, most
companies have a value for the average hourly wage paid
to each type of employee and a generic percentage of their
overall burdened costs (i.e., what is the real total additional cost
of each employee to the company) based on the average wage.
Breaking down these individual inputs to this level of potential
detail and including them individually into a COO calculation
would very rarely be worth the additional effort involved to
answer most questions and make good decisions based on the
purpose of the COO analysis. In practice, some individual cost
factors included in the pyramid may not be included in an
individual COO analysis or a very rough estimate may
be included if the COO modeler has a very good reason to
believe that these cost factors will have no significant impact on
Page |4
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE
the COO analysis. Also, the COO modeler can test their
assumption by performing the COO calculation using a range
of potential values to see if there is any significant impact (i.e.,
a sensitivity analysis). For example, almost all COO analyses
involve performing multiple COO calculations with different
cost factor inputs to compare their results. One common usage
is to compare the COO analyses/values for one supplier's
equipment to another supplier's equipment to better
understand the relative long-term costs of choosing to purchase
one instead of the other. If a particular cost factor/input (e.g.,
burdened labor cost) is not expected to be significantly
different between the choices being modeled, then the absolute
accuracy of these inputs will not have much impact on the
final analysis and decision. Of course, there are several other
important factors in this equipment selection decision besides
COO, but it should be included to make the best overall
decision. Other typical usages are to compare the COOs
before and after a proposed equipment upgrade/modification
or process change to determine the cost impact and return on
investment (ROI).
(*SEMI Standards licenses (company license or single viewer license) for the
standards mentioned above can be purchased at www.semi.org/standards.)
Page |5
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

III. VDMA
Forecasting Model VDMA Standard 34160

This standard sheet describes a forecasting model for calculating lifecycle costs. The model does not include financing, capital
costs or other price effects. The period under consideration is the machine lifecycle starting with the acquisition and ending
after a respective specified utilization period. Anything related to the machine prior to its acquisition or after the period of
consideration is only included in the analysis if it affects costs during the period under consideration. Modifications that are not
performed in the course of maintenance are treated as further utilization and terminate the respective period under
consideration. If lifecycle costs across several modifications are to be analyzed, the model has to be applied repeatedly for the
respective sub-periods.

Figure 4 — Structure of forecasting model for calculating lifecycle costs during period of consideration

Page |6
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

In the case of modifications, the residual value from the period prior to the modification represents the acquisition costs for the
new period under consideration. The model differentiates between three phases: preparatory, operation and further utilization.
For each phase, the individual, relevant cost pools are identified. The lifecycle costs arise from the sum of the costs in the three
phases.

For each phase, the model determines the cost pools along with the pertinent calculation rules. During the preparatory phase,
the focus is on acquisition, start-up and provision of the required infrastructure.

In the operating phase, the model differentiates according to four aspects, "material", "product", "utilization" and "maintaining
functionality", taking into account the production process conditions specified by the customer.

In the further utilization phase, the model includes both costs arising from disposal or refurbishment of the machine as well as
proceeds resulting from a residual value analysis or sale of the machine.

LCC

Preparatory Phase Operating Phase Further Utilization Phase

E B V
Detail Level

E1 E2 E3 IH1 IH2 RK3 SO1 V1 V2 ... V7

...

V1.1 V1.2 V1.3

Flat Rat e
Expansion of Model
IH1.1 IH1.2 ... IH1.6
n ...
...

Figure 5 — Cost pool structure of forecasting model

This standard sheet specifies items, names and detail levels for ordinary costs as the cost pools are structured. Individual cost
items are added on a step-by-step basis, thus resulting in the total cost of a cost pool.

Page |7
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

P Preparation
E1 Acquisition
E2 Infrastructure costs
E3 Other preparatory costs

B Operating costs
IH1 Maintenance and inspection
IH2 Repairs
IH3 Unscheduled repairs
RK1 Occupancy costs
MK1 Material costs
EK1 Energy costs
HB1 Production and process materials
EN1 Disposal costs
PK1 Personnel costs
WK1 Tool costs
RU1 Set-up costs
LK1 Warehouse costs
SO1 Other operating costs

V Further utilization costs

V1 Dismantling
V2 Residual value
V3 Other further utilization costs

Figure 6: Cost elements von VDMA 34160

The structure of the model facilitates a systematic expansion of its components.

 For example, on each aggregate level of a phase, the "Other" item, e.g. SO1, allows the inclusion of additional
costs.

 Each amount can be specified as defined or entered as a lump sum, e.g. flat rate for Set-up Costs RK3.

 Each input value can be described in greater detail on an additional sub-level; e.g. Disassembly V1 could be
further differentiated into personnel expenses and material costs, if required.

 In order to describe various aspects of a cost pool, the respective level can be indexed, which is the case with,
e.g. Maintenance and Inspection IH1.

The forecast is based on the framework conditions for machine operation as specified by the customer or operator. A
comparison of the forecasts for different machines requires the identical framework conditions for the technical specifications
and the specified basic data for the forecasts.

Page |8
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

Technical availability of machines and production lines - VDI 3423

This guideline is intended to specify the necessary definitions, for single machines and system components, and for the entire
system, as well as to list the criteria for a continuous and traceable record of the operational procedure. This is required to:

a) document occupied times,

b) identify downtimes due to organizational or technical problems, or to preventive maintenance
c) determine availability, utilization ratio and failure rates

The data thus determined can be used to:

• locate weak points
• furnish proof in case of warranty claims
• compare different production facilities and their components
• compare different divisions of a company
• calculate economic efficiency
• support investment decisions
• support life cycle cost estimations

Figure 7: Time definition (based on VDI 3423)

Further information about VDMA Standard 34160 and its application for the photovoltaic industry can be found at
http://www.vdma.org/en_GB/article/-/articleview/1180530.

Page |9
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

IV. Comparison VDMA – SEMI

Looking deeper at SEMI and VDMA standards it can be ascertained that armed with a dedicated definition of the various
parameters the different standards can be compared to each other.

To better understand the inherent differences in the standards one needs to look at the viewpoint of the standards. SEMI E10
reflects the timing observation from a user perspective, whereas, the VDI 3426 observes the timing from an equipment
maker’s perspective. This basically can be considered as the main difference between the two standards (as shown in Figure 8.)

Figure 8: Time definitions VDI 3423 – SEMI E10 as given by the standards

From the point of view of the equipment supplier, organizational Down Time, Standby Time, Engineering Time and
unplanned use time are all in the domain of the end-user (as shown in figure 9.)

Figure 9: Responsibilities for Time aspects based on SEMI E79

P a g e | 10
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

It is necessary to be clear which of these Time aspects have to be converted during the purchasing process.

To make the SEMI E10 and VDI 3426 more compatible to each other, the following conclusion between SEMI and VDMA was
reached (see also Figure 9 and 10):

 The organizational Downtime T0 will be set to “zero” for the LCC calculation and Standby Time will be set to
“zero” for the COO calculation. By doing so, the “manufacturing time” equals the “utilization time”.
 The “engineering time” will be added to and considered part of the “productive time.” As a result, the expanded
“productive time” of SEMI E10 equals the “use time” Tn of VDI 3426
 The baseline for all calculations is 7 days/week, 24 hours/day, called Total Time in SEMI E10. As a result, all time
calculations of VDI 3426 are referring to “planned production time” TPRODÄ , while SEMI E10 uses different
denominators depending on the specific metric. If Nonscheduled Time is set to or assumed to be zero, then VDI
3426 Production Time and SEMI E10 Operations Time will be equal.

Figure 10: Time definitions VDI 3423 – SEMI E10 by using this agreement

The related OEE-figures will be the same by using this agreement.

Figure 11: Comparison of OEE Figures

P a g e | 11
 TOTAL COST OF OWNERSHIP IN PV MANUFACTURING GUIDE

V. Conclusion
Strategy-oriented cost management includes detailed life-cycle models, such as total cost of ownership and life-cycle costs.
Standards and standardization belong to the core strategic management toolsets utilized in every industry. Standardization
facilitates the technical and economic cooperation at national, local and international levels. Certainly, this also applies for total
cost of ownership calculations in the PV industry. This Guide illustrates in a simple way how every stakeholder in the PV
industry can either use SEMI standards or VDMA standards to reach the same results. It therefore acts as an instruction to
translate the perspective of a PV manufacturer (SEMI E10) into the perspective of a machine supplier (VDI 3423) and vice
versa. In this context it is decisive for the tool user to carefully adjust the parameters to the same base level. The PV industry is
in a maturing stage. Therefore, it is crucial to consider cost of ownership calculations. This significantly reduces
miscommunication between suppliers and manufacturers.

SEMI and VDMA are committed to support the PV industry in calculation of cost of ownership to better understand the
needs in purchasing discussions.

VI. References
Relevant Standards
 VDI 3423:2011: Availability of machinery and equipment, definitions, meanings, time recording and calculation
 VDMA 34160:2006: Forecast model for the lifecycle costs of machinery and equipment
 SEMI E10: Specification for Definition and Measurement of Equipment Reliability, Availability, and Maintainability (RAM)
and Utilization
 SEMI E35: Guide to Calculate Cost of Ownership (COO) Metrics for Semiconductor Manufacturing Equipment
 SEMI E79: Specification for Definition and Measurement of Equipment Productivity

Authors
Stephan Raithel, SEMI, sraithel@semi.org
James Amano, SEMI, jamano@semi.org
David Bouldin, Fab Consulting, david.bouldin@sbcglobal.net
Dr. Florian Wessendorf, VDMA, florian.wessendorf@vdma.org
Dr. Frank Buenting, VDMA, frank.buenting@vdma.org

About SEMI
SEMI is the global industry association serving the nano- and microelectronic manufacturing supply chains. SEMI supports its
members through a global network of offices and services including advocacy, standards development, industry research, and
events. SEMICON expositions are held in every major semiconductor manufacturing region of the world — the premier
platform for business, networking, collaboration, and education. SEMI maintains offices in Bangalore, Beijing, Berlin, Brussels,
Grenoble, Hsinchu, Moscow, San Jose, Seoul, Shanghai, Singapore, Tokyo, and Washington, D.C. www.semi.org.

About VDMA
VDMA represents over 3,100 mainly small/medium size member companies in the engineering industry, making it one of the
largest and most important industrial associations in Europe. VDMA’s headquarter is in Frankfurt, Germany. VDMA maintains
representative offices in Beijing, Shanghai, Kolkata, Tokyo, Moscow, Sao Paulo, Brussels and Berlin. www.vdma.org.

P a g e | 12